class MadhavDaluka {
  role         = "B.Tech Mathematics & Computing Student"
  education    = "SVNIT Surat, Gujarat, India 🇮🇳"
  expertise    = ["💻 Competitive Programming", "🌐 Web Development", "📊 Data Science & AI"]
  achievements = [
    "🏆 Winner at Capital Odyssey 2026 (SVNIT)",
    "🏆 GirlScript Summer of Code 2026 (Top 50 Contributor)",
    "⚔️ LeetCode Knight (Max Rating 1923)",
    "🥇 Top 12 at CodeCraft (MINDBEND SVNIT)",
    "🎯 Finalist at MetaCode (IIT Kharagpur)",
    "🏅 Rank 5 at ACM Summer Challenge (SVNIT)"
  ]
  certifications = [
    "📊 Certificate of Excellence (Top 10%ile): Summer Analytics 2026 - Consulting & Analytics Club, IIT Guwahati",
    "🎓 Harvard CS50x: Introduction to Computer Science",
    "📈 Kaggle: Pandas, Data Visualization, Intro to ML, Intermediate Machine Learning, Intro to SQL & Advanced SQL",
    "☁️ Cloudinary: Introduction to Media IQ for Developers",
    "🏢 Celonis: Celonis Foundation, Process Mining Fundamentals",
    "🍃 MongoDB: Building AI Agents with MongoDB"
  ]
}

Core Languages

Backend & Database

Data Science & AI

Web & Systems Tools

A high-performance in-memory key-value storage engine built from scratch with custom data structures, key expiry mechanisms, and zero-sleep automated testing.

• Tech Stack: C++17, CMake, Catch2
• Key Highlights:
  • Custom storage engine featuring a separate-chaining hash table, min-heap key expiry (lazy & periodic cleanup), and $O(1)$ LRU eviction using a doubly-linked list.
  • Implemented Redis-style command layer (SET, GET, EXPIRE, TTL, etc.) with 8 automated unit tests using a custom fake clock for deterministic TTL validation.
  • Benchmarked performance on 1M keys against std::unordered_map, reaching within 9% insert throughput (6.57M vs 7.16M ops/s) and lookup speeds within 2%.
  • Analyzed LRU cache efficiency under an 80/20 power-law workload, achieving a 31% hit rate at 10% capacity (vs. 10% uniform) and reaching 72%–98% of optimal cache efficiency.

A data-driven decision support tool that acts as an analytical Head Coach to draft the ultimate playing XI using constrained knapsack problem solving.

• Tech Stack: Python, PuLP, Streamlit, Pandas, Plotly
• Key Highlights:
  • Engineered a linear programming model to solve a constrained knapsack problem, maximizing team fantasy points based on 2025 player stats.
  • Built an interactive Streamlit and Plotly dashboard for dynamic budget tracking and role-based squad constraint enforcement.

A modular, database-driven Discord application featuring dynamic state management and integrated natural language AI.

• Tech Stack: Node.js, Discord.js v14, Gemini API, MongoDB Atlas
• Key Highlights:
  • Built a persistent RPG economy system (inventory, jobs, gambling) utilizing MongoDB Atlas and Mongoose for dynamic state persistence.
  • Integrated asynchronous REST APIs, including Google's Gemini AI for natural language processing and PokéAPI for live data fetching.